Defects of fashion clothing

The proposed solution allows to carry out an automated analysis of defects present on materials (raw materials and semi-finished products) and products (prototypes and final products) through Computer Vision and Video Analysis mechanisms.

Defect detection, such as holes/scratches/trims and non-conforming seams

During the production of the prototype or final product, the manufacturing and packaging process can lead to results that differ from those of the starting model. Computer vision systems based on machine learning will be used to detect the defects.
The Reco solution includes:

1. The development of the lighting system and the acquisition set up;
2. The creation of the training dataset;
3. The implementation of the defect detection subsystem;
4. The implementation of the defect classification subsystem

Color defect detection

The color of the material used for production (leather, fabric, rubber, etc.) may appear different from that of the example model.
The Reco solution includes:

1. The development of a lighting system and the acquisition set up;
2. The analysis of the appropriate color space and chromatic distance;
3. The implementation of an artificial vision system for the matching / detection of chromatic anomalies (handcrafted computer vision technology).

All Reco 3.26 systems can be easily integrated into the customer's pre-existing infrastructure

Implementation

Solution applied for the inspection of clothing for companies in the textile and fashion sector

Background

OVS